Project A3 - Comprehensive simulation framework for modelling complex auditory discrimination experiments
The goal of this project is to provide a software tool to simulate, i.e., virtually perform, speech recognition of listeners with impaired hearing in complex communication environments. For this, a robust automatic speech recognition system is employed that does not require prior information about empirical speech recognition performance.
Based on simulations, the outcome of speech tests and also the optimal behavior (e.g., head position) for the best test result is predicted depending on the used hearing device. The focus lies on understanding which (signal) properties are relevant for robust (human) speech recognition and which part of it can be integrated by listeners with impaired hearing.